| Actually, the thing that makes it "bewildering" rather than "simple"
is that only particular combinations work, in a lot of places.
Take boxes (there are a lot of other examples, this is just one...)
A *simple* message would be maybe 4 "boxes" (desktop, deskside,
server and refrigerator), all of which can have an Intel or Alpha
CPU boards, and, if Alpha, all 3 OSes. Instead, there are all
sorts of "if x then not y", which means the simple message gets
VERY blurred.
Want a low end system? Get an Alpha XL. Ops - you want Unix
or VMS, then you need xxx. Want a fast system? Get an EV5.
But that only runs in a 30K workstation, and, not the same box
where the P6 runs so you can't upgrade. Want a big server?
Get a Turbolaser. Ops - want NT, make that a Sable instead.
Want a portable to run a particular Alpha OS - it had better
be VMS...
(We spend money trying to partition markets to preserve margins,
while our competitors from the PC space are spending money to make
lower and lower cost systems at higher and higher volumes making
our attempts at margin preservation look worse and worse, since
we aren't following the curve down as fast.)
At least in terms of "boxes", we seem to do more when less (variety)
would do. Further, the boxes we do ship (because we have so many
different ones...) frequently aren't at the leading edge because we're
doing so many. Example: We have a low-cost P6 system, but where's
performance and price/performance competitive EV5 system (since,
unlike P6, EV5 has been shipping for the better part of a year...)?
That hardly seems like an agressive, nimble, competitor, nor do
the collection of boxes provide a simple clear message.
|
| the article under discussion... numerous forwards and such
deleted..... posted w/o permission... mods, if inappropriate,
please delete...
InfoWorld Publishing Company
December 3, 1995
Digital Equipment Corp. may yet prove that Jurassic
Park is not the only place where something can be brought back to life.
Regarded as a prime candidate for extinction as the decade opened, the
company appears to be rebounding with a flourish by blending industry-
standard technologies with its proprietary products.
But if the company has any hopes of maintaining that momentum, it needs to
improve on its capability to execute on product strategies, a problem that
has plagued the company for the past five years.
Instead of fighting with industry leaders such as Microsoft Corp. and
Intel Corp., Digital is now focusing on using Windows NT to unify a
seemingly bewildering array of operating system and processor technologies.
The company's strategy is to offer industry-standard PC products that can be
tightly integrated with powerful enterprise systems based on 64-bit versions
of OpenVMS and Digital Unix.
To further that integration, the company plans to announce this week at
the Digital Equipment Computer Users Society conference in San Francisco a
variety of system management tools that work across Windows NT and OpenVMS.
That effort will be further enhanced in the next 12 to 18 months by
bringing the Win32 APIs to OpenVMS via a pact with Bristol Technology Inc.
that will be announced this week.
"I like what I see of their selling NT across hardware platforms as a
unifying software strategy," said John Dunkle, president of Workgroup
Strategic Services Inc., in Portsmouth, N.H. "[Digital] crafted a smooth
upgrade strategy for users that want to move back and forth between Intel
and Alpha systems."
Ultimately, this strategy will enable developers to create Windows
applications that transparently run across a three-tier architecture made up
of Windows clients, Windows NT servers, and OpenVMS in the enterprise.
"We expect NT to have the highest growth rate in that middle or
integration tier," said Enrico Pesatori, vice president and general manager
of Digital's Computer Systems division.
In a Unix Expo speech in October, company Chairman Bob Palmer said he
expects Windows NT and Unix to each account for 40 percent of the server
market by the year 2000.
Long committed to delivering its RISC-based Alpha servers and workstations
with Windows NT, the company is now beginning to make a concerted push with
Windows NT on Intel processors to fulfill this strategy.
This effort has already helped Digital rebound from four years of losses,
with the company posting a profit of $121.8 million for its 1995 fiscal
year, compared to a $2.2 million loss for 1994.
Despite the company's economic turnaround, some observers think its lack
of speed in bringing products to market and an effort to support everything
in a sometimes overwhelming portfolio of technologies -- Windows, Windows
NT, OpenVMS, Digital Unix, OSF, Pentium, and Alpha -- may haunt its comeback
efforts.
"Digital is still most challenged by its lack of speed," said Richard
Buchanan, president of The Buchanan Group, in Peterborough, N.H. "Not just
the general speed of change and unpredictability, but how quickly
competitors can grab the lime-light from them, like Sun [Microsystems Inc.]
is doing with its Internet strategies."
And some observers believe the company's lingering love affair with
OpenVMS may impede the progress it is making in other environments.
"It keeps their installed base happy as they keep evolving things like
OpenVMS and Digital Unix," said one longtime Digital analyst. "But they are
investing in [OpenVMS and Digital Unix] now to the point where it is slowing
down development of other, more important, operating systems."
And although corporate users say they generally like what they hear coming
from Digital in the way of client/server strategies, some say the company's
failure to aggressively embrace Unix System V pushed potential customers
away.
"Like some corporate shops we have already voted with our dollars for
other client/server solutions. [Digital] would have to come in with systems
at least an order of magnitude [better] for us to change our minds," said
Tim McCallister, a systems engineer with a large Chicago-based railroad.
Digital officials argue there will be more than ample opportunity to make
up for lost time in the client/server arena, particularly as IS shops
consider a move to 64-bit platforms and/or transition their Unix and NetWare
applications to Windows NT over the rest of the decade.
Although accused of moving slowly, Digital officials pointed to a number
of leading-edge products and services that might appeal to IS shops that
need the additional power:
* Pentium Pro server and workstation solutions running NT;
* The fastest RISC processor line with the Alpha;
* The first commercially available 64-bit OS implementations; and
* A service and support organization that now generates $7 billion in
revenues.
"We, like a lot of other companies, can provide [customers] with products
that are scalable and portable," said Howard Elias, head of marketing for
Digital's Personal Computer business unit. "But IS also needs a company
that can manage and support client/server strategies across the enterprise,
and I think we have proven [we can do] that."
With PC revenues expected to top $3 billion in fiscal 1996, Digital is
finally able to present a multiplatform hardware strategy to corporate
accounts that carries some weight. For example, Digital rolled out seven
NT- based "personal workstations" in late September with motherboards that
let users plug in either an Alpha or an Intel chip.
That design should translate into sharply reduced manufacturing costs,
Digital officials said.
"We now have the manufacturing economies [of scale] to build a large
workstation business that complements our PC business and can pass those
savings along to users," said Ed Muth, director of planning and product
management for Digital's Windows NT business segment.
That should help move Digital into the top half-dozen or so PC suppliers,
up from No. 11 where it now resides, said Bruce Claflin, manager of the
company's worldwide Personal Computer business unit.
"Digital has already had its catharsis and redefined everything it was
doing. It now recognizes what its core strengths are and how to leverage
them," Claflin said.
Digital might also be able to better take advantage of technological
superiority in other areas, such as notebooks. Early next year, Digital
will unveil its first Pentium HiNote systems, a move that may help it garner
market share from notebook leaders such as IBM and Toshiba America Inc. (See
"Digital breathes new life into HiNote line," Nov. 27, page 1.)
But while Digital can also point to these achievements, the pressure is
far from over.
The company's revolutionary portable design, widely praised when
introduced last year, was unsuitable for use with the higher powered Pentium
processor, and most competitors have already introduced Pentium-powered
notebooks.
Even Digital's most impressive technological feats -- 64-bit Unix and the
Alpha processor -- might be fleeting advantages.
Hewlett-Packard Co., now the No. 2 computer manufacturer in the United
States after IBM, has taken responsibility -- along with The Santa Cruz
Operation Inc. -- for developing an industry-standard 64-bit version of Unix
by 1997, which Digital has promised to support.
Meanwhile, as Digital begins a strong push into the lower end business and
consumer PC markets, it finds HP there a step ahead and already gaining
rapid acceptance.
"Digital is coming back, but the nature of the competitive environment
today won't allow them to ever regain that No. 2 spot from HP," said Frank
Dzubeck, president of Communication Network Architects Inc., a consultancy
in Washington.
If enthusiasm among users for 64-bit solutions continues to build slowly,
Digital's lead in that market may not mean much. It will likely allow HP a
chance to regroup and deliver competitive 64-bit solutions a year or two
from now, when corporate users may be more receptive.
"If Digital can capture most of the early adopters of 64-bit while
continuing with their NT strategy on the low end, they can gain a true
advantage," Buchanan said. "But as is true about most things in life,
timing is everything."
That could be equally true for Digital's Alpha line, easily the most
powerful of the RISC processors currently on the market, but overshadowed in
sales by the PowerPC, Sun's Sparc architecture, and even HP's PA-RISC
architecture.
Although Alpha's sales grew 77 percent during the past year, some
observers suspect the chip's volumes are not high enough to generate
significant profits, and may not be until Digital is able to sign some
licensing deals with large players.
"The volumes aren't there for Alpha to be very profitable yet," said one
analyst. "Right now they only have 25 `design-ins,' all of them with small
players using the chip for rather specialized uses."
Digital hopes to offset that disadvantage partially by building systems
that can use either Pentium or Alpha processors, and by using NT as a common
OS bridging the two hardware platforms. It's a strategy Digital hopes will
lead customers to Alpha when their processing requirements increase.
"Of all the RISC-based platforms out there, we still believe Alpha enjoys
a performance advantage over Intel," Pesatori said.
PC-based systems take off...
U.S. shipments of Digital's portable PCs, desktop PCs,
and PC servers for the past five quarters
Shipment date Units
Third quarter 1994 82,200
Fourth quarter 1994 117,200
First quarter 1995 126,000
Second quarter 1995 115,500
Third quarter 1995 110,200
...Alpha-based systems chug steadily
U.S. shipments of Alpha-based servers and workstations broken out,
and their combined total, for the past five quarters
Shipment date Servers Workstations Total
Third quarter 1994 3,137 8,700 11,837
Fourth quarter 1994 3,626 11,541 15,167
First quarter 1995 7,053 11,196 18,249
Second quarter 1995 7,751 12,330 20,081
Third quarter 1995 7,166 11,415 18,581
|